Catalytic C-H Bond Activation and Knoevenagel Condensation Using Pyridine-2,3-Dicarboxylate-Based Metal-Organic Frameworks.

Three 1D coordination polymers (CPs) [M(pdca)(H2O)2] n (M = Zn, Cd, and Co; 1-3), and a 3D coordination framework {[(CH3)2NH2][CuK(2,3-pdca)(pa)(NO3)2]} n (4) (2,3-pdca = pyridine-2,3-dicarboxylate and pa = picolinic acid), have been synthesized adopting a solvothermal reaction strategy. The CPs have been thoroughly characterized using various spectral techniques, that is, elemental analyses, FT-IR, TGA, DSC, UV/vis, and luminescence. Structural information on 1-4 was obtained by PXRD and X-ray single-crystal analyses, whereas morphological insights were attained through FESEM, AFM, EDX, HRTEM, and BET surface area analyses. Roughness parameters were calculated from AFM analysis, whereas dimensions of small domains and interplanar spacing were defined with the aid of HRTEM. CPs 1-3 are 1D isostructural networks, whereas 4 is a 3D framework. Moreover, 1-4 display moderate luminescence at rt. In addition, 1-4 have been applied as economic and efficient porous catalysts for the Knoevenagel condensation reaction and C-H bond activation under mild conditions with good yields (95-98 and 97-99%), respectively. Notably, 1-3 can be reused up to seven cycles, whereas 4 can be reused up to five catalytic cycles with retained catalytic efficiency. Relative catalytic efficacy toward the Knoevenagel condensation reaction follows in the order 2 > 1 > 3 > 4, whereas 2 > 4 > 1 > 3 for C-H activation. The present result demonstrates synthetic, structural, optical, morphological, and catalytic aspects of 1-4.

Introduction

Development of porous heterogeneous molecular catalysts is a demand of current industrial research. n class="Chemical">Metal–organic frameworks (n>n class="Chemical">MOFs) frequently recognized as porous coordination polymers (PCPs) are materials with excellent crystallinity composed of metal ions/metal clusters with organic linkers.[1−6] Switchable MOFs are a type of smart materials that undergo distinct and reversible structural changes upon exposure to the external stimuli, thus finding interesting technological application.[7] The geometry of a ligand governs crystal topology of MOFs and tunes micro-to-nano crystal morphologies.[8] For instance, the ZIF-MOF family offers excellent chemical and thermal stability and adjustable porous structures.[9] Over past couple of decades, PCPs have drawn tremendous attention and established prodigious worldwide interest not only owing to versatile designability, excellent tunable porosity, structural diversity, miscellaneous topologies, and high surface areas but also due to their fascinating applications in waste water treatment,[10−13] catalysis and gas storage/separation,[14−17] chemical sensing,[18−20] optoelectronic materials,[21−26] heterogeneous catalysis,[27−34] water oxidation,[35,36] energy storage and conversion,[37−40] luminescent materials,[41−44] and so forth. CPs having metal nodes and a variety of bridging organic linkers comprise exposed active metal centers, thereby providing a high degree of metal dispersion for unambiguous catalytic applications.[45,46] In addition, most emerging and advanced applications of crystalline PCP materials are heterogeneous catalysis.[59−67] Because of their multicomponent nature, feasible functionalized catalytic active site PCP materials have been proved as ideal platforms for heterogeneous catalysis and they have been utilized in a wide range of chemical reactions with promising catalytic performance.[56,82,83] CPs have demonstrated excellent morphology-dependent heterogeneous catalytic performance catalytic activities for a wide variety of organic reactions, viz., aldol condensation,[57,58] Henry reaction,[59−61] Michael addition, multicomponent reaction,[62] C–H activation,[47,63] Friedel–Crafts reactions,[64] Tandem reactions,[65] and Knoevenagel condensation.[66−70] Moreover, the structural tunability of PCPs leads to the elegant tailoring of a chemical environment at catalytic sites, thereby causing chemo, regio-, stereo-, and/or enantio-selectivities.[71−73] In addition, the crystalline nature of PCPs provides opportunity to well-disperse active sites at a molecular level, thereby favoring to the mechanistic studies.[48−55]

In partin class="Chemical">cular to the Knoevenagel n class="Chemical">condensation reaction, a variety of heterogeneous catalysts have been used, that is, zeolites,[74−78] clays,[79−83] organic-functionalized molecular sieves, silicate–organic composite materials,[84] and PCPs,[30,85−89] which demonstrated vast significant advantages over homogeneous catalysts. Remarkably, it is noteworthy to mention that PCP-based heterogeneous catalysts are quite limited toward oxidative conversion of benzaldehyde into benzoic anhydride (C–H activation) which is a very advantageous reagent in organic synthesis,[90−93] for instance, silyl ester’s lactonization,[94−96] asymmetric esterifcations,[97] and synthesis of peptides and drugs.[98−100] Employing homogeneous catalysts, Knoevenagel condensation reactions and oxidative conversion of benzaldehyde into benzoic anhydrides involve various difficulties such as low catalyst loading, poor recyclability, tedious work-up process, longtime consumption, and catalyst contamination, whereas heterogeneous catalysts are easily recoverable and reusable and minimize the undesired waste.[101−105]

Additionally, pan class="Chemical">pyridine-dicarboxylaten> and analogous linkers have been meticulously utilized in the construction of a variety of PCPs.[109−119,156] Among pyridine-dicarboxylates, 2,3-pdca ligand has been least utilized toward production of PCPs possibly due to obstruction of its use as a bridging ligand.[118,156] Notably, under hydrothermal/solvothermal conditions (Scheme ),[106−108] 2,3-pdca rarely assumes a unique coordination mode via in situ mono-decarboxylation at the second position, thereby producing nicotinic acid,[156] whereas 3,4-pdca transforms into isonicotinic acid.[119] Remarkably, the present work entails in situ decarboxylation at the third position on 2,3-pdca to produce picolinic acid (pa) under the solvothermal reaction[120] (Chart c and Scheme ). Although 2,3-pdca-containing MOFs/CPs are documented in the literature, however these have seldom been used as heterogeneous catalysts especially in the Knoevenagel condensation reactions and oxidative conversion of benzaldehyde into benzoic anhydrides.[63,121,122]

Scheme 1

In Situ Mono-decarboxylation of 2,3-pdca to Form (Right) Nicotinic Acid (∼Half Dozen Examples) and (Left) Picolinic Acid in the Present Work

Chart 1

Binding Modes of 2,3-pdca in 1–3 (a) and 4 (b,c).

Scheme 2

Synthetic Route for Preparation of 1–4

Results and Discussion

FT-IR Spectral Analyses

The FT-IR spectrum of 1 displayed broad and strong vibration at 3418 cm–1 which may be ascribed to the existence of n class="Chemical">pan class="Chemical">coordinated pan>n class="Chemical">water molecules (Figure S19.1).[124] Furthermore, the presence of the strong peaks at 1667 and 1587 cm–1 may be associated with the characteristic of asymmetric stretching (νas) vibrations due to the −COO– group, whereas the peaks observed at 1442 and 1394 cm–1 are attributed to symmetric stretching (νs) vibration of the −COO– group. Relatively larger separations >200 cm–1 between νas(COO–) and νs(COO–) indicate the monodentate coordination mode of the carboxylate group, whereas smaller separation <200 cm–1 suggests the chelating bidentate coordination mode of the carboxylate group.[125−127] The presence of the corresponding Δν value >200 and <200 cm–1 in the same spectrum indicates both bidentate and monodentate −COO– coordination mode in the 2,3-pdca ligand which has further been authenticated by X-ray single-crystal analyses (vide supra). The other important absorption bands that appear in the IR spectrum of 1 are νC–H (2980 cm–1) and νC=C (1464 cm–1) (Figure S19.1).[128] Peaks appearing in the far-IR region indicate the formation of Zn–N and Zn–O bonds.[129] Likewise, 2 exhibited vibration at 3517 cm–1 owing to the coordinated water molecules along with characteristic vibrations associated with νas and νs −COO– groups at 1661 and 1555 cm–1 and at 1441 and 1379 cm–1, respectively (Figure S19.2).[124,125] Weak vibrations present in the range 425–460 cm–1 may be attributed to Cd–N and Cd–O stretching vibrations.[129] CP 3 shows vibrations at 3441 cm–1 (νH2O), 1583 cm–1 (νas–COO–), and 1369 cm–1 (νas–COO–) (Figure S19.3).[124,125,128,130] On the other hand, the FT-IR spectrum of 4 displayed a vibration at 3238 cm–1 which may be attributed to the −N–H stretching band of quaternary amine (CH3)2NH2+. Furthermore, characteristic νas and νs bands of −COO– groups of 2,3-pdca appear at 1633 and 1592 cm–1 and at 1425 and 1382 cm–1 (Figure S19.4),[130] having Δν value >200 cm–1 which indicates that the carboxylates of 2,3-pdca ligand are linked with Cu(II)/K(I) adopting bidentate and/or bridged bidentate coordination modes.[129] Overall, the FT-IR data are consistent with the structure obtained from SC-XRD analysis.

Powder X-ray Diffraction Study

In order to check the structural identity and phase purity of 1–4, powder X-n class="Chemical">ray diffraction (PXRD) analysis was performed at rt (Figure S20). The PXRD patterns of 1, 2, and 4 closely match with the patterns simulated from their respective X-ray single-crystal diffraction data. On the other hand, the major peak profiles of experimental PXRD of 3 closely match with experimental PXRD patterns of 1 and 2. It indicates that n class="Chemical">CP 3 possesses the structure very similar to those of 1 and 2. Notably, the PXRD pattern of 4 varies from the rest of n>n class="Chemical">CPs 1–3 in terms of the appearance of peaks at a low 2θ angle (2θ = 8.12 corresponds to (2 0 0) reflection), and a higher intensity ratio suggests a different and more porous structural framework of 4. The relation of PXRD data and transmission electron microscopy (TEM) image has also been discussed in the Morphological Studies for 1–4 section (vide supra). The PXRD patterns of 1–4 recovered after catalytic reactions have also been obtained after 5 to 7 catalytic cycles (vide supra). Overall, the PXRD analyses of 1–4 are indicative of bulk sample purity.

Crystal Structure Studies

Structural Description of 1 and 2

Solvothermal reactions of pan class="Chemical">2,3-n class="Chemical">pdca with pan>n class="Chemical">Zn(NO3)2·6H2O and Cd(NO3)2·6H2O in the presence of KOH afforded single crystals of [Zn(2,3-pdca)·(H2O)2] (1) and [Cd(2,3-pdca)(H2O)2] (2), respectively. CPs 1 and 2 are charge neutral 1D polymeric chains comprising fully deprotonated ligands 2,3-pdca2– (Figure c,d). Complexes 1 and 2 assume a quite similar structure wherein the metals (Zn2+/Cd2+) act as a node and 2,3-pdca2– serve as a linker along with two coordinated water molecules on each metal center (Figure a,b). Detailed crystallographic parameters for 1 and 2 have been listed in Table . X-ray single-crystal diffraction analysis reveals that 1 and 2 crystallize in the monoclinic P21/c space group and the asymmetric unit consists of one Zn(II)/Cd(II), one 2,3-pdca2–, and two coordinated H2O [Figure a–d, Table ]. Each dianionic 2,3-pdca2– ligand was linked with three metal centers adopting two distinct coordination modes by N∧O chelating coordination to M(II) from one −COO– and M(II)–O–C–O–M(II) bridging coordination from the other −COO– group [M = Zn(II)/Cd(II); Mode-XII-μ3-(κ4N,O2:O3:O3), Supporting Information]. Two coordination sites have been occupied by two aqua ligands leading to a slightly distorted octahedral geometry about the metal center with symmetry codes: (i) x, y, z + 1; (ii) −x + 1, −y + 1, −z + 1; and (iii) x, y, z – 1 for 1, whereas (i) −x + 1, −y + 1, −z + 1; (ii) x, y, z + 1; and (iii) x, y, z – 1 for 2. Furthermore, carboxyl oxygen atoms are involved in intra- and intermolecular hydrogen-bonding interactions.

Figure 1

(a,b) Asymmetric unit of 1 and 2; (c,d) two types of cavities (square and ellipsoid shape) differing in dimensions along the crystallographic “c”-axis. Symmetry code for 1 (i) x, y, z + 1; (ii) −x + 1, −y + 1, −z + 1; and (iii) x, y, z – 1; for 2, (i) −x + 1, −y + 1, −z + 1; (ii) x, y, z + 1; and (iii) x, y, z – 1.

Table 1

Crystal Data and Structure Refinements for 1, 2, and 4

compound	1	2	4
formula	C7H7NO6Zn	C14H14Cd2N2O12	C14H12CuKN3O8
formula weight	266.53	627.09	452.91
crystal system	monoclinic	monoclinic	orthorhombic
space group	P121/c1	P121/c1	Pnma
a (Å)	7.7081(4)	7.8597(5)	21.7641(13)
b (Å)	15.6681(7)	15.9234(10)	17.4903(10)
c (Å)	7.7483(4)	8.0981(5)	9.3685(5)
α (deg)	90	90	90
β (deg)	113.824(2)	114.849(2)	90
γ (deg)	90	90	90
V (Å3)	856.04(8)	919.67(10)	3566.2(4)
Z	4	18	8
Dcalc (g cm–3)	2.0679	2.2644	1.687
μ (mm–1)	2.879	2.383	1.51
F(000)	537.5656	605.1382	1832
θmin, θmax (deg)	2.87, 30.55	2.77, 30.57	25.7, 2.3
hmin–max	–11, 10	–11, 11	–26, 26
kmin–max	–22, 22	–22, 22	–21, 21
lmin–max	–10, 11	–11, 11	–11, 11
reflections collected	2611	2824	3503
data/restraints/parameters	2611/0/163	2824/0/62	3503/594/286
R1, wR2 [I > 2σ(I)]a	0.026866/	0.037138/	0.119709
R1, wR2 (all data)a	0.037753/0.094958	0.041446/0.127273	0.1638/0.2423
no. unique data	2611	2824	3503
no. Observed	2241	2585	2320
no. Variables	163	62	286
Rint	0.0561	0.0388	0.229
wR	0.094958	0.127273	0.2423
GOF on F2	0.760475	1.078909	1.094
(Δρ)max,min (e/Å3)	0.742389(−0.645688)	2.328405(−4.14180)	2.04(−1.15)
(Δ/δ)max, (Δ/δ)mean CCDC	0.0003, 0.0000	0.0292, 0.0013	0.039, 0.000

(a,b) Asymmetric unit of 1 and 2; (c,d) two types of cavities (square and ellipsoid shape) differing in dimensions along the crystallographic “c”-axis. Symmetry n class="Chemical">pan class="Chemical">code for 1 (i) x, y, z + 1; (ii) −x + 1, −y + 1, −z + 1; and (iii) x, y, z – 1; for 2, (i) −x + 1, −y + 1, −z + 1; (ii) x, y, z + 1; and (iii) x, y, z – 1.

One of the pan class="Chemical">con>ordinated water molecules occupies an apical position with a longer Zn–O/Cd–O distance of 2.1651 Å/2.335 Å, whereas the other one assumes an equatorial position with a shorter Zn–O/Cd–O distance of 2.0065 Å/2.226 Å. One oxygen atom from −COO– and one pyridyl nitrogen atom from 2,3-pdca ligand chelate to one Zn(II)/Cd(II) metal having Zn–N/Cd–N and Zn–O/Cd–O bond distances 2.104 Å/2.313 Å and 2.056 Å/2.248 Å, respectively. The transoid angles around Zn(II)/Cd(II) lie in the range of 173.61–176.00°/169.24–176.44°, whereas cisoid angles range from 77.74–99.76° to 72.89–90.77° in 1 and 2 (Tables S13.1.3 and S13.2.2). Two types of M(II)···M(II) distances occur in a double-stranded 1D chain of 1/2; (i) the distance between M(II)···M(II) separated by bridging oxygen of the −COO– is uniformly 3.444 Å/3.629 Å and (ii) the M(II)···M(II) separated by pdca2– ligands is 6.433 Å/6.453 Å (M = Zn, Cd; Schemes and 3, Figures b and 2b). The structure of 1/2 reveals the presence of two sorts of cavities in their polymeric chain; square shape cavities having a dimension of 3.44 × 2.77 × 5.039 Å3/3.629 × 2.989 × 5.171 Å3 are created by two nearest M(II) centers separated by V-shaped bridged oxygens. The dimension has been measured by considering M(II)···M(II) and O···O distances. Ellipsoidal cavities of dimension (6.433 × 6.193 × 4.724 Å3/6.453 × 6.929 × 4.836 Å3 in 1/2) are formed between M(II)···M(II) centers separated by 2,3-pdca2– in the double-stranded chains and the dimension has been calculated by distances between M(II)···M(II) and O···O atoms, centroid–centroid distance of the aromatic rings of bridged 2,3-pdca2– ligands. In earlier reported CPs, Zn(II)···Zn(II) centers separated by bridging oxygen of the −COO– is uniformly 3.294 Å and (ii) the Zn(II)···Zn(II) separated by pdca2– ligands is 8.479 Å.

Scheme 3

Metal–Metal Distances in Two Types of Cavities Present in the Polymeric Double-Stranded Chain of 1 and 2

Figure 2

(a) Asymmetric unit of 4 and (b) dimeric unit of 4. Symmetry codes: (i) −x + 1, −y + 1, −z + 2; (ii) x, −y + 3/2, z; (iii) −x + 1, −y + 1, −z + 1; (iv) −x + 1, y + 1/2, −z + 1; (v) x + 1/2, y, −z + 1/2; (vi) −x + 1/2, −y + 1, z – 1/2; (vii) −x + 1/2, y + 1/2, z – 1/2; (viii) x, −y + 1/2, z; (ix) x – 1/2, y, −z + 1/2; and (x) −x + 1/2, −y + 1, z + 1/2. (c) Cap-stick view of a cage-shaped 3D cavity present in 4 along the crystallographic a-axis. (d) Demonstration of two Cu···Cu distances between adjacent molecules of helical and ellipsoidal cavities along crystallographic-b axes (hydrogen atoms are omitted for clarity).

Structural Description of 4

The molecular structure of 4 was determined by its X-n class="Chemical">ray single-crystal analysis (Figure ). Crystallographic parameters and selected bond lengths and bond angles for 4 have been listed in Tables and S13.3.1–S13.3.3, respectively. n class="Chemical">MOF 4 crystallizes in the orthorhombic crystal system with the n>n class="Chemical">Pnma space group. The solvothermal reaction of 2,3-pdca with Cu(II) in the presence of KOH followed by recrystallization in dimethyl formamide (DMF) leads to the formation of a 3D network {[(CH3)2NH2][CuK(2,3-pdca)(pa)(NO3)2]} (4) via mono-decarboxylation from one unit of 2,3-pdca to form picolinic acid (PA–). By ignoring the disorder, an artificial lowering of the bond lengths with disordered atoms and high values of geometrical parameters has been observed. The counter cation (CH3)2NH2+ is present in the cage to neutralize the negative charge present in the overall complex 4. The (CH3)2NH2+ entity resulted from the decomposition of DMF under solvothermal conditions which is known in the literature.[131] Incorporation of (CH3)2NH2+ as a counter cation in an anionic coordination network has previously been reported, although its source was not mentioned therein.[132−134] Interestingly, the single-crystal structure of 4 revealed an unprecedented in situ decarboxylation of one carboxylate group from 2,3-pdca2– to produce picolinic acid (HPA). However, mono-decarboxylation occurred only from one unit of coordinated 2,3-pdca2– out of its two units linked with metal ions (Figure b,d). The Cu(II) centers assume distorted square pyramidal geometry completed by N2O3 coordination sites offered by 2,3-pdca2– and PA– ligands wherein 2,3-pdca2– adopts two distinct coordination environments; (i) one N∧O chelating site which also bridges through carboxylate oxygen and (ii) terminal coordination through 3-carboxylate. At the same time, PA– coordinated through the N∧O chelating mode to the Cu(II) center. The apical Cu–O bond distance of 2.246 Å is slightly longer relative to the basal Cu–O bonds. The Cu–N and Cu–O bond distances in the basal plane lie in the ranges of 1.966–1.978 and 1.957–2.245 Å, respectively, which are comparable to the analogous systems.[131,135−139] Carboxylate groups and the pyridine ring are almost coplanar in 2,3-pdca. The adjacent distances between Cu(II)···Cu(II) separated by square and ellipsoid shape cavities in the 1D polymeric chain are 3.798 and 6.408 Å, respectively. A local coordination environment with about seven coordinated K(I) centers can be best described as a distorted tetragonal antiprism which consists of two 2,3-pdca2–, two PA–, and one nitrate ligand. The coordination mode adopted by the 2,3-pdca2– ligand in 4 can be classified as Mode μ4-(κ7N,O2:O2,O2′:O2′,O3:O3′).

(a) Asymmetric unit of 4 and (b) dimeric unit of 4. Symmetry pan class="Chemical">con>des: (i) −x + 1, −y + 1, −z + 2; (ii) x, −y + 3/2, z; (iii) −x + 1, −y + 1, −z + 1; (iv) −x + 1, y + 1/2, −z + 1; (v) x + 1/2, y, −z + 1/2; (vi) −x + 1/2, −y + 1, z – 1/2; (vii) −x + 1/2, y + 1/2, z – 1/2; (viii) x, −y + 1/2, z; (ix) x – 1/2, y, −z + 1/2; and (x) −x + 1/2, −y + 1, z + 1/2. (c) Cap-stick view of a cage-shaped 3D cavity present in 4 along the crystallographic a-axis. (d) Demonstration of two Cu···Cu distances pan class="Gene">between adjacent molecules of helical and ellipsoidal cavities along crystallographic-b axes (pan class="Chemical">hydrogen atoms are omitted for clarity).

The pan class="Chemical">polymern>ic 3D framework of 4 wherein 2,3-pdca displays an unprecedented μ7-coordination mode further extends through intermolecular contacts. Another PA– ligand assumes coordination mode Mode-XXV μ2-(κ3N,O2:O2′). The K1–O distances are lying in the range of 2.670–2.941 Å, whereas K2–O distances range from 2.633 to 3.330 Å. Several ellipsoidal cavities with dimensions (7.437 × 7.830 × 11.872 Å3), (7.437 × 7.830 × 11.474 Å3), (3.829 × 5.694 × 11.872 Å3), (3.829 × 5.694 × 11.474 Å3), (11.872 × 11.474 × 9.940 Å3), (6.408 × 6.078 × 4.496 Å3), and (6.408 × 6.078 × 9.714 Å3) are present in 4 which have been measured by diagonal distances between centroids and centers separated by 2,3-pdca2–/PA– ligands (Figures and 3, S12d and S14). The Cu(II)···Cu(II) centers are separated by two ways with distances 3.798 and 6.408 Å. The apparent 3D cages present in 4 with said dimensions are occupied by (CH3)2NH2+ cations through weak bonding interactions.

Figure 3

Cap-stick view of 4 along the crystallographic a-axis.

Supramolecular Interactions in 1, 2, and 4

The weak interactions through H···bonding, C–H···π, and π···π interactions lead to intern class="Chemical">connectivity of double-stranded chains in 1 and 2, fabricating 2D layers in the bc-plane. Intermolecular O–H···O n>n class="Chemical">hydrogen-bonding interactions between oxygen atoms of water molecule and carboxylate groups lead to the construction of 2D repeated double-stranded chains along the crystallographic b-axis (Figures S2 and S7b). The distance between Zn···Zn/Cd···Cd centers separated by intermolecular hydrogen bonding is 5.393/5.430 Å. In 1, the inter- and intramolecular H···bonding distances lie in the range of 2.666–2.744 and 2.616–2.932 Å, respectively, whereas in 2, intermolecular H···bonding distances lie in the range of 2.663–2.744 Å. Notably, intramolecular hydrogen bonding is not observed in 2. Intermolecular C–H···π interactions form zigzag infinite parallel chains with hexagonal cavities both in 1 and 2 lead to a 2D network along the crystallographic b-axis having a dimension of 3.857 × 4.177 × 6.804 Å3 measured between O(2)···O(2), O(4)···O(4), and O(7)···O(7) in 1 (Figures S2–S4) and a dimension of 4.338 × 4.482 × 7.064 Å3 measured between O(3)···O(3), O(5)···O(5), and O(1)···O(1) in 2 (Figures S9–S11). In spite of a similar structural framework in 1 and 2, the intramolecular π···π stacking distances are lightly different [3.220, 3.243, and 3.373 Å in 1 and 3.354 and 3.391 Å in 2] while none of these entails intermolecular π···π stacking interactions. The centroid–centroid distances between two parallel aromatic rings of bridging pdca2– are 4.725 and 4.836 Å, respectively, in 1 and 2, whereas centroid–centroid distances between intra- and interchain six-membered aromatic rings are 4.724 and 5.372 Å in 1 and 4.836 and 5.486 Å in 2. The nearest Zn···Zn separation in 1 is 3.444, whereas in 2, the adjacent Cd···Cd distance is 3.629 Å (Scheme ). Slightly longer distances in 2 may be due to larger ionic radii of Cd(II) in comparison with Zn(II). Distances between μO(1)···μO(1) in 1 and 2 are 2.777 and 2.989 Å, respectively. Notably, intra-/intermolecular H···bonding interactions involve all oxygen atoms of 1 and 2 which contribute significantly toward stability of intermolecular chains (Table ). Prominently, these compounds contain both the Lewis basic (the noncoordinated oxygen atoms) and acidic centers (coordinated Zn(II)/Cd(II) ions) and coordinated water molecules as the potentially good leaving group in the same unit that enables these systems to find potential application as a bifunctional catalyst (Figure ). On the other hand, supramolecular assemblies in 4 through intermolecular N–H···O (2.827 Å) weak interactions between 2,3-pdca2– carboxylate oxygen and N–H of the dimethyl ammonium cation lead to 3D extensive chains (Figure S18). The intermolecular C–H···π distances were observed in the range of 2.473–2.879 Å (Figure S13), whereas intermolecular π···π stacking interaction distances occurring between pyridine rings of 2,3-pdca2– are observed to be 3.343 Å (Figure S17).

Table 2

Selected Hydrogen Bond Geometry (Å) in 1, 2, and 4

D–H···A	D···H	H···A	D···A	∠DHA
1
O7-Hc···O6	0.82(3)	1.82(3)	2.6149(18)	165(3)
O7-Hd···O6	0.82(3)	1.92(3)	2.7451(19)	177(3)
O4-Ha···O5	0.75(3)	2.03(3)	2.7749(19)	173(3)
O4-Hb···O2	0.82(3)	1.85(3)	2.6658(17)	173(3)
2
O5-Ha···O4	0.8700	2.03(2)	2.753(3)	140(3)
O1-H1a···O6	0.8700	1.895(9)	2.748(3)	166(3)
O1-H1b···O6	0.8700	1.913(7)	2.668(3)	144.4(9)
4
N1S-H1Sd···O6	0.92	1.939(14)	2.830(14)	162.4(4)
N1S-H1Se···O6	0.92	1.939(15)	2.830(14)	162.4(4)

Figure 4

(a) Cap-stick view of the 2D double-stranded chains incorporating dimeric 1 between two layers. (b) Inter- and intramolecular H···bonding interaction in 2 measured from the crystallographic a-axis. (c) 3D view of 4, resulting from O–H···O hydrogen bonding with a dimeric complex along the crystallographic b-axis.

(a) Cap-stick view of the 2D double-stranded chains inn class="Chemical">pan class="Chemical">corporating dimeric 1 class="Chemical">n>n class="Gene">between two layers. (b) Inter- and intramolecular H···bonding interaction in 2 measured from the crystallographic a-axis. (c) 3D view of 4, resulting from O–H···O hydrogen bonding with a dimeric complex along the crystallographic b-axis.

Morphological Studies for 1–4

The morphological information was acquired for straightforward n class="Chemical">pan class="Chemical">compan>rison of surface structures of CPs/MOF 1–4 so as to rationalize their distinct optical and catalytic behavior. Shape and microstructure of the resulting 1–4 have been investigated via scanning electron microscopy (SEM) (Figure top; Figures S25–S28). To have deeper insights, SEM, TEM, and atomic force microscopy (AFM) analyses for 1–4 have been comprehensively illustrated. SEM, TEM, and AFM analyses revealed fairly appealing and idiosyncratic external morphological behavior by 1–4 (Figures and 6). SEM images of 1–4 display homogeneous nanocrystallites aggregated with an excellent pore size in the clusters. CPs 1–4 are shown as crystalline materials through SEM images with an elongated block shape and particle sizes of about 30, 3, and 15 μm for 1–3, respectively. The morphological view of 1 displays interconnected crystalline fluffy (feathery) layered sheets (Figure , top 1), whereas SEM images of 2 show elongated unruffled nanorods with an approximate individual length ranging from 300 to 500 nm having uniform diameters of 20–40 nm. Moreover, SEM images of 3 exhibit little bent uniform layer with obvious small pores. Notably, 2 displayed an irregularly layered fibrous array with a larger pore size relative to that of 1 (Figure , top 2). In sharp contrast, SEM analysis of 3D MOF 4 exhibited flower-like porous morphology with granular clusters (Figure , top 4). The hierarchical structures in 1–4 are bound to possess a larger specific surface area capable of absorbing substrate(s) or facilitating heterogeneous catalytic activity to enable these materials as catalysts. In addition, the EDX analyses reveal the presence of C, N, O, and M with the respective weight ratio of 35.4, 10.3, 31.2, and 23.1 (M = Zn, 1), 37.2, 16.4, 39.5, and 6.9 (M = Cd, 2), and 32.8, 8.9, 41.4, and 16.9 (M = Co, 3), whereas C, N, O, K, and Cu elements with the respective weight ratio of 29.3, 37.2, 29.5, 7.8, and 26.2 (4) (Figures S29–S32). Overall, the EDX analysis strongly supports successful synthesis of 1–4.

Figure 5

SEM (top) and HRTEM (bottom) images of 1–4.

Figure 6

AFM images of 1 (a,e), 2 (b,f), 3 (c,g), and 4 (d,h).

In addition, TEM analysis of thin films of 1 and 2 revealed typical flat sheet morphologies and apparently visual cavities (Figure bottom 1 and 2). Interestingly, despite structural resemblance, the TEM images exhibited different porous surfaces of 1 and 2 along with n class="Chemical">considerable liaisons among the arrays. The HRTEM image of 1 exhibited the interplanar d-span>cing of 0.705 and 0.392 nm which pan>n class="Chemical">corresponds to the respective (1 0 0) and (0 4 0) lattice plane of a monoclinic unit cell of [Zn(pdca)·(H2O)2]. The origin of these fringes in the HRTEM images is clearly related to the PXRD data of 1. In the PXRD analysis of 1 (Figure bottom 1), major diffraction peaks at 2θ values are 11.286, 12.543, 16.904, 17.757, 21.113, 21.152, 21.798, 22.683, 23.937, 25.969, 27.530, 31.209, 32.577, 33.023, 34.313, 34.518, 34.593, 35.484, 36.669, 36.702, 38.144, 40.075, 42.234, 42.311, and 43.456 which may be assigned to the reflection of (0 2 0), (1 0 0), (−1 2 0), (−1 2 1), (0 −3 1), (−1 3 0), (1 −1 1), (0 4 0), (1 −2 1), (0 −4 1), (−2 0 2), (−1 5 0), (−1 −4 2), (1 −1 2), (0 6 0), (1 −2 2), (2 −2 1), (−3 1 1), (−1 6 0), (−1 −2 3), (−3 2 2), (2 −4 1), (−3 1 3), (0 −7 1), and (−3 2 3) crystal planes (Table S14.1). Likewise, the HRTEM image of 2 displayed crystalline nanorod-shaped crystals and exhibited fringes having the interplanar d-spacing of 0.367 nm corresponding to the (0 0 2) lattice plane of a monoclinic unit cell of [Cd(pdca)·(H2O)2] (Figure , bottom 2). The positions and relative PXRD peaks of 2 at around 2θ = 11.104, 12.401, 13.259, 13.593, 17.253, 20.845, 24.204, 25.426, 26.505, 27.785, 29.545, 30.721, 31.712, 32.018, 33.717, 35.050, 35.950, 36.367, 37.100, 37.845, 38.419, 39.488, 40.639, 41.859, and 43.62 which may be apportioned to the reflection of (0 2 0), (1 0 0), (0 −1 1), (−1 1 0), (−1 2 1), (−1 3 0), (0 0 2), (−0 −4 1), (−2 0 2), (−1.–3 2), (0 −3 2), (−1 5 0), (1 0 2), (−2 4 1), (−2 4 0), (1 −5 1), (0 −6 1), (−1 6 1), (−3 2 2), (−2 5 0), (−0 −2 3), (2 −4 1), (−3 1 3), (−3 2 3), and (2 −2 2) crystalline lattice planes, respectively (Table S14.2). The CP 3 comprises small particle and crystalline morphologies in its HRTEM images which shows an interplanar d-spacing of 0.4068 nm consistent to the (1 1 1) lattice plane of a monoclinic system (Figure bottom 3). Positions and the relative PXRD diffraction peaks of 3 are more or less matching with those of 1 and 2. The positions and the corresponding PXRD peaks of 3 at about 2θ = 12.187, 13.366, 14.038, 14.687, 17.061, 17.437, 20.425, 21.895, 23.155, 23.254, 24.392, 24.515, 25.576, 29.363, 29.392, 29.622, 30.361, 30.438, 30.517, 30.921, 33.446, 34.260, 34.358, 35.271, 35.294, 35.588, 37.139, 37.242, 38.539, 38.828, 38.894, 39.016, 39.195, 39.776, 39.819, 39.837, 40.521, 40.585, 41.129, 41.365, 41.538, 42.261, 42.263, 43.481, 43.595, 43.738, 43.790, and 43.803 may be ascribed to the reflection of (−1 0 1), (1 0 1), (0 −1 1), (−1 1 0), (0 0 2), (1 −1 1), (0 −1 2), (−1 −1 2), (−2 1 1), (1 −1 2), (−1 2 0), (−2 0 2), (−1 2 1), (−3 0 1), (−1 −2 2), (−2 2 0), (−2 2 1), (1 −2 2), (1 −1 3), (3 0 1), (−2 2 2), (−3 1 2), (0 −2 3), (−1 3 0), (2 −2 2), (2 −1 3), (−3 0 3), (−3 2 1), (1 −1 4), (4 0 0), (−3 1 3), (−1 −3 2), (−2 3 0), (−2 3 1), (−2 −1 4), (1 −3 2), (−4 1 0), (2 −3 1), (−4 0 2), (2 0 4), (−2 3 2), (4 −1 1), (1 −2 4), (−1 −3 3), (−1 0 5), (2 −3 2), and (−3 2 3). Notably, the HRTEM image of 4 shows a more compact surface which may be ascribed to the lattice (CH3)2NH2+ cations present in the cavities. HRTEM images of 4 exhibit the interplanar d-spacing of 0.386 nm which corresponds to the (2 2 2) lattice plane of an orthorhombic unit cell of {[(CH3)2NH2][CuK(2,3-pdca)(pa)(NO3)2]} (Figure bottom 4). The PXRD pattern of the 4 having 2θ = 8.118, 9.564, 10.107, 10.704, 11.451, 12.977, 14.429, 16.278, 18.845, 20.293, 20.627, 21.249, 21.503, 23.019, 23.079, 24.521, 24.702, 24.725, 24.79, 25.065, 25.584, 25.596, 27.19, 27.261, 27.32, 27.502, 27.832, 27.89, 27.959, 29.435, 29.74, 29.80, 29.894, 30.185, 30.263, 30.483, 30.529, 30.537, 30.645, 30.651, 30.702, 32.033, 32.175, 32.401, 32.461, 32.499, 34.643, 36.134, 36.855, 37.472, 38.35, 41.265, 41.959, 41.963, 42.596, 45.129, 48.779, and 49.715 which may be attributed to the reflection of (2 0 0), (2 1 0), (0 2 0), (0 1 1), (1 1 1), (2 2 0), (1 2 1), (4 0 0), (4 0 1), (0 4 0), (2 0 2), (2 1 2), (0 2 2), (2 2 2), (5 1 1), (6 0 0), (1 3 2), (5 2 1), (3 2 2), (4 0 2), (4 1 2), (3 4 1), (0 5 1), (5 3 1), (3 3 2), (1 5 1), (4 4 1), (0 4 2), (5 0 2), (4 3 2), (2 0 3), (5 2 2), (3 5 1), (2 1 3), (7 0 1), (5 4 1), (6 3 1), (3 4 2), (0 6 0), (1 2 3), (7 1 1), (6 4 0), (1 5 2), (1 6 1), (4 4 2), (0 3 3), (4 2 3), (7 2 2), (5 2 3), (6 4 2), (6 1 3), (7 1 3), (5 6 2), (4 0 4), (1 6 3), (5 2 4), (5 4 4), and (4 5 4) lattice planes of the nanocrystallites confirm well with its structure (Table S14.4).

AFM studies were also pursued to quantify the minimum and maximum values of precipitate size and accumulation n class="Chemical">pattern of 1–4. Momentarily, the surface roughness of the samples was characterized by AFM and exn class="Chemical">amined following the existing literature procedure.[140−142] The AFM image of 1 reveals uniformly packed polycrystals with a size of the crystals in the range of ∼30–50 μm (Figure a,e). Moreover, the micelle-like morphology specifies that 1 has a tendency to self-aggregate and further assemble into the crystalline structures. The CP 2 exhibits polymeric, spherical, granular, and porous domains which are clearly visible in the AFM image. These porous and granular structures may provide an increased surface area in 2 for catalytic applications (Figure b,f). Furthermore, the AFM image of 3 exhibits the formation of a well-established crystalline material which gets converted into twisted ciliated fiber-like morphology spread over the film (Figure c,g). The AFM image of 4 demonstrates roughly porous and shaggy layered morphologies having a very small uniform granule-like surface (Figure d,h). The 2D and 3D maps of 1–4 are also given in Figure which indicate different surface morphologies. Average roughness (Ra) obtained from the films of 1–4 is 0.28, 2.52, 0.32, and 2.06 nm, respectively. Moreover, section analysis of the 2D images across the line in 1–4 affords the average tube diameter of around 12–15 nm (Figures S37–S41). The root mean square (rms) roughness (Rq) has been calculated to be 0.332, 2.71, 0.328, and 2.17 nm for 1–4, respectively (Figure S41). The low-to-high values in rms roughness are probably ascribed to the noncovalent intermolecular interactions on the surfaces of the polymers 1–4.[143−146] As Rq is more sensitive to the peak and valley relative to that of Ra, typically Rq is 3 to 15% higher than Ra in these materials. A smaller roughness is associated with greater density and finer precipitates, whereas larger precipitates tend to increase the roughness value (Ra = 0.28–2.52 nm in 1–4). Overall, the SEM, TEM, and AFM analyses strongly suggest diverse structural morphology and porous surface of 1–4. The different intermolecular interactions may be responsible for morphological disparities in 1–4.

UV/Vis and Luminescent Spectral Analyses

Electronic absorption spectra for the suspended particles of 1–4 have been ren class="Chemical">corded in n>n class="Chemical">DMF (Figure S22). CPs 1 and 2 comprising diamagnetic Zn(II) and Cd(II) metals exhibited broad absorption bands at 286 and 287 nm, whereas 3 and 4 involving paramagnetic Co(II) and Cu(II) metals displayed bands having the absorption maximum around 289 nm. Therefore, observed absorption bands for 1–4 may be assigned to the ligand-centered transitions. It has been well-established that d10-transition metal complexes reveal fascinating luminescence properties.[147,148] Hence, the solid-state photoluminescent behavior of 1–4 was also investigated at rt. CPs 1 and 2 encompassing d10-metal centers and conjugated organic ligands were expected to serve as facilitating inorganic–organic hybrids for prospective application as chemical sensors and in photochemistry.[149−153] Notably, 1–4 exhibited moderate luminescence in the solid state.[154] CPs 1–3 showed emission spectra having maxima at 398, 390, and 385 nm along with a shoulder peak at 421, 414, and 422 nm, respectively, upon excitation at their respective wavelengths (286–289 nm), whereas 4 displayed an emission band at 387 nm (λex, 289) along with a shoulder peak at 424 nm, respectively (Figure S23). The emission of 1–4 may be ascribed to the ligand-centered transitions and/or ligand-to-metal charge-transfer transitions.[118,155−160]

TGA and DSC Analysis of 1–4

To further assess the thermal stability of 1–4, thermogravimetric analysis (n class="Chemical">pan class="Gene">TGA) was performed on crystalline samples from 40 to 800 °C under a pan>n class="Chemical">nitrogen atmosphere at a heating rate of 10 °C/min (Figures a and S21). A weight loss of 8.49% until 132 °C followed by another weight loss of 7.37% until 228 °C was observed for 1 which corresponds to the release of trapped and coordinated solvent molecules (calculated 13.54%). The loss of coordinated water molecules in 2 was completed by 169 °C with a weight loss of 9.38% (calculated 11.48%). In 3, a weight loss of 14.23% was occurred between 40 and 133 °C which is consistent with the loss of coordinated water molecules (calculated 13.84%). The rest of the framework remains thermally stable up to ∼300 °C; thereafter, the desolvated framework starts decomposing the organic components and completes by ∼460, ∼417, and ∼430 °C for 1–3, respectively. Compounds 1–3 exhibit similar decomposition processes which corroborate well with their structures (SCXRD and PXRD). Notably, the thermogravimetric histogram of the 4 does not exhibit any inflexion point before ∼216 °C, which indicates its good thermal stability and strongly supports the lack of water molecule(s) [crystalliferous, lattice, or coordinated (Figures a and S21)]. Exceeding 220 °C, it was sharply decomposed by losing almost entire organic 2,3-pdca2– ligands with a weight loss (wt loss) of 53.99%. The remaining dark-brown scorched compound (unidentified) is obtained at 800 °C, which may be the metal oxides or metal carbonates. Overall, TGA analysis indicates good thermal stability of 1–3 and excellent thermal stability of 4.

Figure 7

TGA (a) and DSC (b) profile of 1–4.

pan class="Gene">TGAn> (a) and DSC (b) profile of 1–4.

In addition, DSC curves for 1–3 exhibited endothermic peaks followed by exothermic peaks (n class="Chemical">pan class="Chemical">collapse/desolvation) of the framework (Figures b and S22). The values of ΔHf (J/g) and ΔSf were calculated from DSC endothermic peaks (class="Chemical">n>n class="Chemical">DTGmax = 128 °C; ΔH = +181.2 J/g, ΔSf = 1.42 and DTGmax = 210.8 °C; ΔH = +185.8; ΔSf = 0.88; DTGmax = 320 °C; ΔH = 21.36 J/g, ΔSf = 0.067) for the first stage while an exothermic peak (DTGmax = 378.18, 552.49 °C, ΔHm = −1687 J/g, ΔSf = −3.63) for the second stage for 1. In contrast, 2 displays an endothermic peak (DTGmax = 165 °C; ΔHf = +154.9 J/g; ΔSf = 0.94) in the first stage and an exothermic peak (DTGmax = 388.93, 569.13 °C, ΔHm = −4118 J/g; ΔSf = −8.59) in the second stage. On the other hand, ΔHf and ΔSf for 3 have been calculated from endothermic peaks (DTGmax = 174.9, 366.7, and 519 °C; ΔH = 27.85, 65.83, and 36.63 J/g; ΔSf = 0.16, 0.18, and 0.071) in the first stage and (DTGmax = 337.80, 452.55, and 542.58 °C, ΔHm −730.6 J/g, ΔSf = −1.64) in the second stage. Additionally, 4 displayed a sharp endothermic peak at 257.7 °C followed by breakdown of the framework. The ΔHf and ΔSf for 4 were calculated from an endothermic peak (DTGmax = 257.7 °C; ΔH = +198.6 J/g, ΔSf = 0.77) for the first stage while in the second step, DTGmax = 401.16 and 564.59 °C, ΔH = −3015 J/g; ΔSf = −6.24. Overall, the DSC studies suggest a better stability of 4 over 1–3, whereas 1–3 follow the stability order 3 > 2 > 1 (Figures b and S22). The greater stability of 4 over 1–3 is also associated with the absence of coordinated water molecules in 4. Moreover, intermolecular associations may be the major factor of stability order 3 > 2 > 1.

While pan class="Chemical">con>rrelating DSC studies with the crystal structure, it can be stated that the framework structure converts from stable six-coordinated distorted Oh to less-stable four-coordinated Td geometry upon removal of two water molecules; hence, the process is endothermic in 1–3. Eventually, the remaining dark-brown scorched compound (unidentified) obtained at 800 °C may be the more-stable metal oxides or metal carbonates; hence, this process is exothermic. On the other hand, DSC of 4 shows one endothermic peak at 257.7 °C which may be attributed to the conversion of more-stable to less-stable species upon removal of lattice solvent/guest molecule followed by exothermic breakdown of the framework to attain metal oxides/carbonates.

Catalytic Studies on 1–4

The lability of ligands propan class="Chemical">viden>s coordinately unsaturated active catalytic centers to carry out various catalytic reactions. The Knoevenagel condensation and C–H activation reactions are one of the most studied heterogeneous catalytic reactions performed for porous and framework materials of various metal centers, that is, Co, Cu, Zn, Ag, Cd, In, and so forth. The PCPs connected to several flexible ligands containing distinct aromatic rings are appropriate scaffolds for such catalytic reactions because metal centers exhibit variable oxidation states and coordination numbers.[161] The Knoevenagel reaction is well-known not only as a weak base-catalyzed model reaction but also as a C–C bond formation reaction.[89,162−166] In the present work, catalytic synthesis of 2-benzylidenemalononitrile has been chosen mainly because these intermediates are often used in polymers, perfumes, cosmetics, fine chemicals, pharmaceuticals, and drugs.[141,182,185−191]

Activation Method

Catalysts 1–4 were heated in an oven at 100 °C for 24 h under vacuum and a n class="Chemical">pan class="Chemical">N2 atmosphere to remove the air/solvent molecules from the pores of n>n class="Chemical">CPs. The heating temperature was selected considering thermal analyses so as to avoid structural rupture.

Knoevenagel Condensation Reaction Catalyzed by 1–4

The optimized activated catalysts 1–4 (5.0 mol %) were added in the reaction mixture of pan class="Chemical">benzaldehyden> (1.0 mmol) and malononitrile (1.1 mmol) and the resulting mixture was stirred at rt for a different time scale depending upon progress of the reaction monitored by thin layer chromatography (TLC). Information on the product formation was analyzed using 1H and 13C NMR and FT-IR techniques (Figures S42–S46). Upon reaction completion, the mixture was diluted with 5 mL of CH3OH to dissolve the organic products which was followed by centrifugation cum filtration and the supernatant liquid was evaporated to dryness. The pure product was obtained by recrystallization in MeOH. The catalyst was removed by filtration and washed with EtOAc which was recovered, dried, and reused as required (Scheme ). 2-Benzylidenemalononitrile (Table , entry 1) mp: 83.5 °C; 1H NMR (500 MHz, CDCl3; δ, ppm): 7.89 (d, J = 8.0 Hz, 2H), 7.78 (s, 1H), 7.63 (t, J = 7.5 Hz, 1H), 7.54 (t, J = 7.8 Hz, 2H). 13C NMR (125 MHz, CDCl3; δ, ppm): 160.03, 134.76, 131.01, 130.83, 129.68, 113.83, 112.67, 82.93. FT-IR (KBr, cm–1): 3426, 3033, 2956, 2867, 2777, 1695, 1668, 1592, 1569, 1491, 1461, 1375, 1336, 1300, 1231, 1202, 1182, 1163, 1144, 1101, 1001, 972, 776, 756.

Scheme 4

Synthesis of 2-Benzylidenemalononitrile Using Heterogeneous Catalysts 1–4

Optimization of the Catalyst and Solvent for the Knoevenagel Condensation Reaction

The objective of these optimizations was to develop epan class="Chemical">con>nomical and most effective catalyst and solvent system for the Knoevenagel reaction. To seek for the possibility of heterogeneous catalytic behavior of 1–4, the one-pot Knoevenagel reaction was carried out in different optimized solvents. At the onset, catalysts 1–4 were separately employed in different sets of the model Knoevenagel reaction using n>n class="Chemical">benzaldehyde and malononitrile reactants to yield 2-benzylidenemalononitrile at rt under solvent-free conditions. The reaction process can be constantly monitored by TLC, illustrating that this catalytic reaction indicated a significant amount of product after 25 min. Taking a random amount of catalysts for the fixed timescale of 25 min, 1–4 provided the product yields 61, 68, 59, and 53%, respectively. Notably, the said reaction completes in 4 h under catalyst-free conditions. To select the most appropriate reaction condition, the parameters such as catalyst amount and best solvent were optimized (Tables S1–S4). The performance of 1–4 was examined by varying the catalytic amount to 1, 2, 3, 4, 5, and 10 mol % in the reaction and observed that increasing the catalytic loading from 1 to 5 mol % leads to the enhanced yield of the product; however, escalating the amount from 5 to 10 mol % could not significantly increase the product yield at the same time scale. The kinetics of conversion profiles show that the initial reaction rates and the final product yield were higher with 5 mol % relative to the other catalytic loadings; hence, the best catalytic loading of 1–4 chosen was 5 mol % (Figure S47, Tables S1–S4, entry 6). In addition, the reaction was investigated with various solvents, that is, H2O, EtOH, CH2Cl2, CH3CN, toluene, and benzene, and also under solvent-free conditions wherein moderate yields were observed using these solvents (Tables S1–S4, entries 8–14), whereas solvent-free conditions afforded excellent yields of the product in the 25 min time scale. The lower yields in H2O and EtOH relative to that of the solvent-free reaction may be attributed to the solvation of the active functional groups by these solvents as well as due to hydrogen-bonding interaction between active protonic sites of 1–4, thereby diminishing in the catalytic efficiency. Therefore, 5 mol % catalytic loading of 1–4 under the solvent-free conditions at rt was designated as the optimum condition (Tables S1–S4).

Among the tested solvents, pan class="Chemical">ethanoln> has been observed as the best solvent for the Knovenegal reaction; hence, further experiments for the optimization of catalyst loading were performed in ethanol. A blank-controlled experiment afforded 38% conversion of benzaldehyde after 4 h in ethanol at rt, whereas 95–98% conversion could be accomplished using catalysts 1–4 under the same conditions and 25 min time scale (Figures S52 and S53). The time-conversion plot exhibiting the comparative reaction rate in the presence and absence of 1–4 clearly suggests that the reaction is promoted only in the presence of catalysts (Figure S53). Furthermore, leaching experiments have also been conducted under identical conditions to assure whether the reaction is stimulated by the heterogeneous catalysts not due to the active sites leached into the solution. Therefore, the reaction between benzaldehyde and malononitrile was initiated in the presence of 1–4 under identical conditions and the solid catalyst was removed by filtration after 10 min while the resulting solution was allowed to continue up to 25 min (Figure S55). Notably, the reaction rate was significantly reduced in the absence of a catalyst which indicates that the reaction is exclusively catalyzed by 1–4. Comparison of the reaction rates for 1–4 clearly indicates a higher initial reaction rate for 2 relative to those of 1, 3, and 4 which may be attributed to either the lack of diffusion restrictions or possibility of increasing coordination number for the Cd(II) center. Furthermore, recovered catalysts have also been analyzed by PXRD which revealed the structural and morphological stability of 1–4 (Figure ).

Figure 8

PXRD patterns of 1–4 freshly prepared, recovered after the first cycle, and recycled after seven (1–3) and five (4) cycles.

PXRD patterns of 1–4 freshly pren class="Chemical">pared, repan class="Chemical">covered after the first cycle, and recycled after seven (1–3) and five (4) cycles.

Hence, the stability of 1–4 was expan class="Chemical">aminen>d by recycling the catalyst in the successive runs up to seven cycles and observed consistent performance for 1–3; however, 4 displayed consistent performance up to five cycles only (Figure S50). Relatively lower reusability observed for 4 may be attributed to the presence of K(I) centers along with Cu(II) centers in the framework. Moreover, the stability of these catalysts during the reusability experiments was also ascertained by comparing the PXRD pattern of the fresh with those of recovered catalysts up to the respective consistent performance (Figure ). The catalytic efficiency of a heterogeneous catalyst with respect to the previously reported catalysts can be best assessed through calculation of turnover number or turnover frequency; therefore, 1–4 have been exhaustively compared with earlier reports. However, the yield achieved in the presence of 1–4 at rt is preferable over the catalysts which operate either at 60–130 °C or involve large time scale reactions (12–24 h).[23a] High yield (98%), ambient temperature (rt), and shorter time (4 h) along with catalyst stability up to many cycles demonstrate the improvement of catalysts 1–4 relative to the earlier reports (Table S11).[23b,62] The salient features of 1–4 comprise high performance in the Knovenagel reaction under solvent-free conditions, functioning under mild reaction conditions, short reaction time, wide substrate scope, and high stability.

Plausible Reaction Mechanism

pan class="Chemical">CPsn> can act as Lewis bases or acids in the catalytic medium depending on the nature of the ligands and metal ions as well as the coordination environment. In 1–4, metal centers Zn(II)/Cd(II)/Co(II)/Cu(II) impart Lewis acidic sites and the carboxylate group present in the 2,3-pdca linker can act as a weak Bronsted base. Thus, 1–4 can be considered as a potential acid–base catalyst for the Knoevenagel condensation reaction in which acidic and basic sites can synergistically catalyze the reaction to improve its efficiency. As observed in the SC-XRD data, the offset packing of the networks in 1–3 perhaps blocks the channels supporting the nonporous nature and hence, the condensation reaction may be occurring on the catalyst surface. On the basis of crystal structure and based on the previous reports,[75,167] a plausible mechanism for the Knoevenagel reaction on the Lewis acidic site has been illustrated (Scheme ). The reaction is plausibly initiated by the attack of the polarized carbonyl oxygen from the aldehyde with the metal center of the respective CP, followed by subsequent opening of the weakly coordinated chelating carboxylate oxygen from the Lewis acid site. Simultaneously, detachment of an acidic proton of the methylene group of the malononitrile produces a carbonium anion. In the next step, the carbonium anion reacts with the carbonyl group of benzaldehyde to give an intermediate which undergoes rearrangement and elimination of water to afford the final product.

Scheme 5

Proposed Mechanism for the Knoevenagel Reaction, Catalyzed by 1–4

Recyclability of Catalysts 1–4

Catalysts 1–4 might have propan class="Chemical">viden>d lodgings for activated guests in their channels, consequently, the prominent Knoevenagel condensation reaction was selectively stimulated to afford the product in good yield. To examine the recyclability of 1–4 upon reaction completion, the solid catalysts were collected and separated dissolving the reaction mixture in ethanol followed by filtering and drying to reuse for another set of the same reaction. After filtration and drying, the catalysts were subjected to a vacuum oven at 80 °C for 4–6 h to remove adsorbed guests. The vacuum-dried catalyst was then reused for the next cycle of the same reaction. This experiment was repeated seven times and found that 1–3 perform quite well, whereas 4 starts decreasing in its catalytic activity after five catalytic cycles (Figure ). The deactivation of the catalysts after the seventh and fifth cycle may be attributed to the pore blocking by the generated product.[178,179] Remarkably, PXRD patterns of the recycled CPs 1–4 after the seventh and fifth runs, respectively, exhibited peak profiles similar to those of the fresh compounds (Figure ). Relative performances of 1–4 with some previously reported catalysts for the Knoevenagel condensation suggested that our catalysts are comparative in terms of cost-effectiveness, ambient temperature, and short time scale of the reaction (Tables S11 and 3).[33,88,168,180,181]

Table 3

Knoevenagel Reaction between Malononitrile and Benzaldehyde Catalyzed by 1–4 in Ethanol

entry	catalyst (mol %)	optimized catalyst quantity	time (m)	yield (%)
1	[Zn(pdca)·(H2O)2]	5 mol %	5	47
			10	67
			15	82
			20	92
			25	97
2	[Cd(pdca)·(H2O)2]	5 mol %	5	59
			10	83
			15	92
			20	95
			25	98
3	[Co(pdca)·(H2O)2]	5 mol %	5	53
			10	69
			15	83
			20	93
			25	96
4	[(CH3)2NH2][CuK(pdca)(PA)(NO3)2]	5 mol %	5	51
			10	74
			15	85
			20	92
			25	95

Hot-Filtration Experiments

To pan class="Chemical">con>nfirm the heterogeneous nature of the catalysts, hot-filtration experiments were also performed.[169−174] In this context, the solid catalysts 1–4 were removed from a hot solution by filtration for 10 min after initiating the catalytic run. The reaction of the filtrate was then monitored for another 15 min wherein insignificant catalytic conversion was observed (Figure S55). It indicates almost no leaching of metal from the catalyst and thus the heterogeneous nature of the catalysts. The filtrate was also analyzed by atomic absorption spectroscopy (AAS), which indicated a very low concentration of free metal(II) ions (0.000017–0.000079%) from catalysts 1–4 that leached out into the reaction solution (Table S9).[175−177]

Synthesis of Benzoic Anhydrides via Aldehydic C–H Activation Using Catalysts 1–4

To a solution of pan class="Chemical">benzaldehyden> (1.0 mmol) and catalyst 1/2/3/4 (5 mol %) in CH3CN (2 mL), a 70% solution of TBHP (1.5 equiv) in CH3CN was added gradually over 10 min under stirring. Furthermore, the reaction temperature was increased to 70 °C followed by stirring for another 1 h under a N2 atmosphere and the progress of the reaction was monitored by TLC (Scheme ). After reaction completion, the solvent was evaporated under reduced pressure and the residue was purified using silica gel column chromatography (hexane/EtOAc, 4.5:0.5). Benzoic anhydride: Colorless crystalline powder; yield: 98%; Anal. Calcd for C14H10O3 (226.23): C, 74.33; H, 4.46; N, 0.0. Found: C, 74.36; H, 4.48, N, 0.0. R = 0.5797, 1H NMR (500 MHz, CDCl3): δ 8.13 (d, J = 7.5 Hz, 2H), 7.63 (t, J = 7.5 Hz, 1H), 7.50 (t, J = 7.8 Hz, 2H). 13C NMR (125 MHz, CDCl3): δ 162.48, 134.66, 130.67, 128.99, 128.59. IR (KBr, cm–1): ν = 3438, 3065, 3036, 3011, 1789, 1725, 1686, 1650, 1599, 1493, 1452, 1345, 1314, 1279, 1173, 1098, 1076, 1049, 996, 870, 801, 778, 702.

Scheme 6

Syntheses of Anhydrides via Aldehydic C–H Bond Activation Using Heterogeneous Catalysts 1–4

To employ 1–4 as heterogeneous catalysts for the production of pan class="Chemical">benzoic anhydridesn> via C–H bond activation, the best reaction pan class="Chemical">conditions and various other parameters such as solvent, oxidant, amount of catalyst, and reaction time were optimized.[182,183]

Optimizations for C–H Bond Activation Reactions in the Presence of 1–4

The effect of various solvents such as pan class="Chemical">Cn class="Chemical">H2Cl2, class="Chemical">n>n class="Chemical">ClCH2CH2Cl, chloroform, ethanol, water, diethyl ether, THF, DMF, and CH3CN was investigated independently for the C–H activation of benzaldehyde in the presence of the catalyst and TBHP, as a model reaction (Table ). There are definite evidences on solvent-dependent C–H bond activation of benzaldehyde. The results of the present study revealed the maximum yield obtained using a CH3CN solvent; therefore, it was chosen as an ideal solvent for this reaction (Tables and 4).

Table 4

Solvent Optimizations for the C–H Bond Activation Reaction of Benzaldehyde Using 1–4a

entry	solvent/medium of the reaction	temp (°C)	conversion (%)	selectivity (%)
1	CH3OH		18	45
2	C2H5OH		19	48
3	CH2Cl2		trace	52
4	C2H4Cl2		5	58
4	DMF		23	64
5	THF		12	72
6	H2O		24	78
7	benzene	rt	29	35
8	toluene		35	93
9	diethyl ether		70	96
10	CH3CN		89	98

Reaction circumstances: benzaldehyde (1.0 mmol), TBHP (1.5 equiv), catalyst (5 mol %), and solvent (2 mL).

Reaction circumstances: n class="Chemical">pan class="Chemical">benzaldehyde (1.0 mmol), class="Chemical">n>n class="Chemical">TBHP (1.5 equiv), catalyst (5 mol %), and solvent (2 mL).

Optimization of the Oxidant

Among oxidants such as air, pan class="Chemical">TBHPn>, H2O2, and oxone, TBHP was optimized as the best oxidant for the abovementioned reaction (Table , entry 7). It is worth mentioning that using H2O2 as an oxidant leads to the conversion of a large part of aldehyde to carboxylic acid and the negligible desired product was observed.

Table 5

Screening of Reaction Conditions and Optimization of the Best Oxidanta

entry	oxidant used for desired product	% age yield
1	air/oxygen	N.O.
2	CHP	N.O.
3	oxone	N.O.
4	K2S2O8	N.O.
5	H2O2	16
6	TEMPO	N.O.
7	TBHP	83

Reaction conditions: benzaldehyde (1 mmol), oxidant (1.5 equiv), and CH3CN (2 mL). N.O. = not observed.

Reaction pan class="Chemical">con>nditions: pan class="Chemical">benzaldehyde (1 mmol), oxidant (1.5 equiv), and pan class="Chemical">CH3CN (2 mL). N.O. = not observed.

Effect of the Catalyst Quantity

The amount of the catalysts 1–4 was optimized in pan class="Chemical">CH3Cn class="Chemical">N in the presence of class="Chemical">n>n class="Chemical">TBHP as an oxidant. The best results obtained using 5 mol % of catalysts, whereas the exceeding catalytic quantity does not show any noticeable enhancement in the product yield (Table ).

Table 6

Optimization of Catalytic Amount of 1–4 in the Synthesis of Benzoic Anhydride in CH3CN at rta

entry	catalyst	amount of catalyst (mol %)	(%) yield
1.1	1	1	43
1.2		2	54
1.3		3	65
1.4		4	73
1.5		5	97
1.6		10	97
2.1	2	1	58
2.2		2	69
2.3		3	77
2.4		4	84
2.5		5	98
2.6		10	98
3.1	3	1	39
3.2		2	46
3.3		3	52
3.4		4	78
3.5		5	96
3.6		10	96
4.1	4	1	61
4.2		2	69
4.3		3	73
4.4		4	76
4.5		5	95
4.6		10	96

Reaction conditions: benzaldehyde (1.0 mmol), TBHP (1.5 equiv), and CH3CN (2 mL).

Reaction pan class="Chemical">con>nditions: benzaldehyde (1.0 mmol), TBHP (1.5 equiv), and CH3CN (2 mL).

After optimization of the model reaction, generality of the reaction was exn class="Chemical">pan class="Chemical">amined using pan>n class="Chemical">benzaldehyde with both electron-rich and electron-poor substituents and the results showed high catalyst proficiency for C–H bond activation cum formation of anhydrides. Reportedly, the aldehydes with a NO2 substituent at para- or ortho-positions in the presence of CuO nanoparticles resulted in the NO-product formation.[27] Other literature reports illustrate the reaction of benzaldehyde in the presence of CuO, CuCl, Cu-MOFs, CuCl2, and TBHP which yielded 35–75% of the anhydrides, after 3 h[63,184] (Table S12, entries 6–8 and 10). Notably, in the presence of 1–4, aldehydes having electron-withdrawing groups lead to the formation of the corresponding anhydrides in relatively good yields, whereas aldehydes with electron-donating groups were converted to the respective anhydrides with good-to-excellent yields (Table S12 entry 13–16). After reaction completion, the catalysts were feasibly separated by filtration and recycled several times. On the other hand, utmost of the reported methods used an equivalent amount of reagents which are expensive and environmentally alarming (Table S12, entry 1–5).

Catalyst Recycling of 1–4

Reusability with pan class="Chemical">con>nsistent performance is one of the important concerns in heterogeneous catalysts. Notably, 1–4 could be recovered simply by filtration, washing with methanol, and drying under a vacuum oven @ 80 °C for 4–6 h and these recovered 1–4 were used for the same C–H bond activation reaction up to four cycles. However, the results indicate that 1–4 can be used for several cycles with minimal loss of activity which is also evident from unchanged HRTEM structural morphology even after seven catalytic cycles (Figure S58). The HRTEM images strongly suggested preserved crystallinity as well as morphology even after seven catalytic cycles of 1–3 and five catalytic cycles of 4. Likewise, no framework degradation of the hydrolysis reaction[63] could be significantly observed in the PXRD patterns (Table ).

Table 7

Synthesis of Benzoic Anhydrides via the C–H Activation of Benzaldehyde Catalyzed by 1–4 (5 mol %) in CH3CN

entry	catalyst	time (min)	yield (%)
1	[Zn(pdca)·(H2O)2]	2	38
		5	56
		10	77
		15	89
		20	97
2	[Cd(pdca)·(H2O)2]	2	42
		5	61
		10	87
		15	94
		20	99
3	[Co(pdca)·(H2O)2]	2	36
		5	51
		10	68
		15	92
		20	97
4	[(CH3)2NH2][CuK(pdca)(PA)(NO3)2]	2	41
		5	53
		10	71
		15	93
		20	98.5

Hot-Filtration Test

Furthermore, leaching of pan class="Chemical">metaln> ions from the heterogeneous surface of catalysts 1–4 was examined through the hot-filtration test. Hence, the reaction was initiated under the optimized reaction conditions and the reaction mixture was filtered off to separate the solid catalysts after 5 min under hot conditions. The filtrate in the absence of solid was continued up to 20 min (Table S10). The analysis of the reaction mixture indicated that the yield of the reaction was insignificantly improved (Figure S56). It strongly indicated the heterogeneous nature of the catalysts 1–4. Moreover, the filtrates were also analyzed by AAS which exhibited a complete absence of free metal(II) ions (0.000000–0.000000) indicating no leaching from catalysts 1–4 (Table S10).

It is noteworthy to mention that the catalytic performance for the Knoevenagel pan class="Chemical">con>ndensation reaction follows the order of 2 > 1 > 3 > 4. The literature reveals that many factors including the particle size, shape, chemical composition, metal–support interaction, and metal–reactant/solvent interaction can have significant influences on the catalytic properties of the catalysts. It is worth stating that since all the catalysts comprise the same pdca2– ligand and 1–3 are isostructural, the difference in product yield is small (95–99%) which may be attributable to the feasibility of basic sites, ionic-covalent nature of the metal–oxygen bonds, ready production of “M–O2 Lewis acid–base” pair, and so forth. All the metal centers [Zn(II), Cd(II), and Co(II)] are six-coordinated and possess entirely the same coordination environment by pdca2– and aqua ligands; therefore, small variation in catalytic performance may be attributed to the size of the metal ions and their electronic configuration. The ionic radii order for these metal ions follows order Cd(II) > Zn(II) > Co(II) which is analogous to their catalytic performances. Therefore, better catalytic performance of 2 over 1, 3, and 4 may be ascribed to the larger size of Cd(II) and its possibility to extend coordination no up to 7-, 8-, and 9-coordination. On the other hand, the catalytic efficiency order follows 2 > 4 > 1 > 3 for the C–H activation reaction which may be attributed to the rms roughness (Rq) which has been calculated to be 0.332, 2.71, 0.328, and 2.17 nm for 1–4, respectively. In addition, five-coordinated Cu(II) in 4 may offer a better binding site for the substrate over six-coordinated metal centers in 1–3. Moreover, the catalytic activities of 1–4 may be ascribed to the combined effect of the metal atom involved in the respective CP and/or porosity of the framework. The CPs 1–4 contain both Lewis acid (Zn, Cd, Co, and Cu center) and Lewis basic (carboxylate and free pyridyl groups) sites, which makes them suitable for bifunctional catalysis.

Experimental Section

Chemicals and Reagents

All the chemicals including pan class="Chemical">Znn>(NO3)2·6H2O, Cd(NO3)2·4H2O, Co(NO3)2·6H2O, Cu(NO3)2·3H2O ≥ 99.00%, and potassium hydroxide (KOH) were procured from HiMedia Chemicals Pvt. Ltd., India. Rare chemical pyridine-2,3-dicarboxylic acid or quinolinic acid was purchased from Sigma-Aldrich Pvt. Ltd., USA. All the solvents such as ethanol and distilled water were purchased from Merck Life Science Pvt Ltd (India). The reagents were procured from commercial sources and used as received. The chemicals and reagents were pure, whereas the solvents were dried and distilled following standard literature procedures.[123]

Conclusions

Four pan class="Chemical">CPsn> [M(pdca)·(H2O)2] [M = Zn(II), Cd(II), Co(II)] (1–3), and [(CH3)2NH2][CuK(pdca)(pa)(NO3)2] (4) have been synthesized implementing a solvothermal reaction strategy. CPs 1–4 have been thoroughly characterized using various spectral techniques, that is, elemental analyses and thermal and spectral techniques. Moreover, the structural information on 1–4 have been obtained by PXRD and X-ray single-crystal analyses, whereas morphological insights were acquired by FESEM, AFM, EDX, and HRTEM analyses and BET surface area analysis. The roughness parameter was calculated from AFM, whereas HRTEM aided in defining the dimensions of small domains and in measuring the interplanar spacing in the CPs. The structurally similar 1–3 are 1D coordination frameworks, whereas 4 is a 3D network. In addition, all 1–4 display good luminescence at rt. Furthermore, feasibly prepared 1–4 serve as efficient and economic porous heterogeneous catalysts for the Knoevenagel condensation and C–H bond activation reaction under mild conditions. Using catalysts 1–4 in Knoevenagel condensation and in C–H activation reaction leads to the product yields in the range of 95–98 and 97–99%, respectively. Notably, catalytic performance of 1–4 was also reconnoitered through reusability and percolating experiments and observed that 1–3 can be reused up to seven cycles with almost consistent catalytic efficiency, whereas 4 retains its catalytic efficacy up to five cycles. The relative catalytic efficiency of 1–4 toward the Knoevenagel condensation reaction has been observed in the order 2 > 1 > 3 > 4, whereas the efficiency for C–H activation follows the order 2 > 4 > 1 > 3. Overall, the present result demonstrates synthetic, structural, morphological, optical, and catalytic aspects of CPs 1–4.